AFM-7 · TRUST BLOWBACK
DESKRUNE
For when you come back.
The pattern: you sent an AI-laundered email to a friend, partner, or coworker. Their reply was different than usual. Cooler. Shorter. They sensed something was off. Now you're rebuilding a relationship the email shortcut quietly damaged.
AI doesn't write emails. It writes objects that look like emails. The people who know your voice can tell. The cost is ambient — they don't say 'this was AI-written,' they just trust you slightly less the next time. Untraceable, expensive.
Hook by hook. If you recognize even one of these in the last week, this is the failure mode that's costing you.
The Deskrune Protocol for AFM-7 is the 'Human-in-the-Loop Pause.' Before sending any AI-assisted communication regarding sensitive topics, you must execute three steps. First, the 'Read Aloud Test': speak the draft; if your tongue trips over the phrasing or it sounds like a press release, rewrite the opening and closing sentences manually. Second, the 'Specific Injection': force yourself to add one sentence referencing a specific, non-public detail known only to you and the recipient. Third, the 'Time-Decay Wait': if the message carries emotional weight, draft it, close the window, wait fifteen minutes, and then review. If the warmth is gone, the send is blocked. This protocol ensures the tool serves the relationship, not the other way around.
Every AFM hits four people. The ADHD adult living it. The partner watching it. The coach explaining it. The boss pricing it. Every kit speaks to all four — because the buyer is sometimes one of the people watching.
You felt the pit in your stomach the moment you saw their cold reply to your AI-drafted message.
You watch them spiral after an email exchange that looked fine on the surface but felt wrong to them.
You see a client losing ground on relationships because they are outsourcing their emotional labor to algorithms.
You notice a top performer suddenly sounding robotic and detached in team communications, signaling a deeper disconnect.
AFMs cluster. If this one hits, the next two probably also do.
Most ADHD-AI loops feel random until you can see them next to each other. Log a moment in the form below — intensity, time, AI tool, brief context. After three or more entries, you'll see your own pattern in the recap. Stays in your browser. Nothing is sent to us.
You don't have to be precise. The point is to capture enough that the pattern becomes visible. Future-you reading this in three weeks will have something concrete to work with.
Deskrune. For when you come back.